Excite@Home To Change Routing Priorities For $$

chrisd writes "I found this one the SJ Mercury news web site, go and read it. The article in essence says that excite expects to make a lot of money from content providers for giving them "special" access to their networks. This is a troubling problem in my mind because, in essence it means that the web will be better for some sites rather than others based on the ability/desire of sites to pay excite." (Read More)

" I am expecially troubled by this because it would mean that sites and companies that have gobs of cash could in essence pay excite to guarantee a better experience than a competitor. It's not clear from the article, but for excite to specifically give higher priority to one than that would by default reduce the priority of competitors.

The question then becomes, how long until someone can literally pay to make it impossible to reach a competitors site. Say how would cnn reader feel if they could no longer reach cnn , but msnbc always came up super fast?

It's a surprise that the same company that appears to be for open access for cable lines would take this approach to cash for network routing. Excite is definately doing the right thing through it's support of open access (imho), but a very wrong thing in pursuing this line of business. "

Comments

These concerns may be unwarranted

[orpheus]

What is the controversy here? According to the article, "those companies had agreed to pay an undisclosed amount per megabit per second in order to plug into the high-speed network." In other words, they are selling bandwidth. If anything, it is less troubling than the many 'preferred vendor' arrangements that have been on commercial networks (e.g. Compuserve, AOL, Prodigy) since the 80's.

By buying bandwidth directly to a network segment, these providers will get better throughput. If you look at the buyers (Akamai, iBeam and Microcast) in the article, you'd see that they defintely have an interest in eliminating netlag and other delays to cablemodem users (who can make best use of their services). I expect other 'wide pipe' providers to follow suit, and consider it both prudent, and a service to all customers. (The revenue stream is welcome, too: Excite@Home lost $1.5 billion on revenues of $337 mill last year. How long do you think they can afford to keep supplying service at current prices at his rate?)

There is a huge distinction between *providing* service* and *denying* it.

You might as well argue that high-bandwidth users are 'crowding everyone else off the Internet' (which has been argued). Howver, this doesn't have that nice conspiratorial anti-business ring, does it?

Nothing wrong with that

[CyberBry]

I don't see anything at all wrong with this, and as an @Home user, I'm very happy with it. There is also a tremendous advantage here for large web sites.

Face it - @Home is massive. With well over a million people with multimegabit per second connections, you can believe that @Home generates a LOT of traffic (I did the math once, and they have over 5Gbps of peering alone). Now, let's say for arguements sake that at any given time, 30Mbps is moving between Yahoo! and @Home. When Yahoo! realizes that they can in essense, colocate a server within @Home, and save that extra 30Mbps, it may be much cheaper to do that than to use up their external peering.

There is an advantage for @Home as well - not only do they save money on peering since they need fewer external links, but also now @Home users have access to this content over much faster connections.

I see how you are saying that companies are paying for performance, but I don't see how that's any different from the current situation. Some companies can afford an OC-3 and can send 300KB/s to an @Home user, while some companies can afford a T1 and can send 20KB/s to an @Home user. Nothing's changing. @Home's peering will not cease to exist when this new service is being offered, so people who cannot afford to place servers on @Home will still be able to send content to @Home users at the same speed as they can now. Since @Home's peering is actually quite nice to most networks (most notable exception is UUNet), their end of the connection can handle as fast as yours can send it.

Re:Nothing new....

[chrisd]

The only difference here is that they are doing this with content providers specifically and not peering with other isps and backbone providers.

Chris DiBona
VA Linux Systems

--
Grant Chair, Linux Int.
Pres, SVLUG

Nope! Read the article

[BranMan]

I don't think this has the evil portents - I read the article and what I get is this:

Everyone on Excite has equal access to all sites (through the regular internet backbones) with some latency and bandwidth determined by that pipe.

Excite is getting companies to pay them for the privalege of hooking high-speed pipes directly from their servers to Excite's routers.

Request for them don't go to the internet backbone - they are routed over these new, direct lines. Faster performance to those sites (less latency, more data/sec., etc.)

Noone is given priority over another. No competitor's service is reduced - in fact it would also be enhanced, since some of the traffic is not going to that common backbone (at least a little).

Put away the pitchforks and torches for now.

Monny-grubbing...

[technos]

Both their push for 'open-access' and 'pay-us-for-routing' are nothing more than trying to protect their bottom line. Their contract with several of the cable providers is up soon, and there is little incentive to go with them again.

This smells rather funny though. No ISP has tried this in the past, so there must be a rather good reason. I'm guessing everyone is afraid when the 'competition' starts screaming to Congress, Congress will hand down far more 'Thou shalt nots' than are needed to rectify the situation.
Congress has shown historically that it not like this sort of thing amongst the telephony and satelite data providers; Why should it like it when e@h does it?

The result? e@h makes a quick buck, and those that paid for 5 years of 'special' routing are going to lose it in a round of regulation.. e@h and the honest ISPs alike are going to have to deal with additional regulation laden on by every SIG waiting for the first ISP regulation test bill to show it's face..

There is absolutely nothing new here

[phoneboy]

ISPs have been doing this amongst themselves for a long time. Companies like above.net do this every day, albeit they host the servers on their site. On the surface, it simply looks like they're making their routing tables more efficient for customers willing to pay for the privilege. Certain companies that make caching servers employ similiar techniques for streaming media (i.e. cache those who pay the piper for the privilege).

Since it gets the content "closer" to the end user's (albeit of @Home), and doesn't negatively affect everyone else out there, I can't necessarily see this as a bad thing.

-- PhoneBoy

Peering Groups and Content Caching

[Kagato]

Considering that Excite, like most national ISPs, are investing in content caching systems I doubt the tool booth is going to be a big deal.

CNN and MSNBC will load off the cache server just as fast. When it comes to streaming audio and video then you're going to see issues. Maybe Real won't work as fast as MS.

On the other hand this isn't anything all that new. PSInet has been offering peering arrangements for some time. And ISP can have a direct connection to PSInet sites, and PSInet has a direct connection to the ISP. Depending on traffic considerations it may not be that bad of a deal. Then again PSInet hosts a lot of servers and T's. Excite really doesn't.

I think this is a pretty big non-issue. Excite is obviously trying to make a dime off high bandwidth video streaming. Somehow I doubt the Opensoure, GNU, FreeWare, Freedom of speech market is going to be hurt. Maybe one porno site will load quicker than another. More power to them.